282152 - Tables not rendering correctly

Status: RESOLVED WONTFIX

(Core :: DOM: HTML Parser, defect)

Description

[hmikron]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.5) Gecko/20041107 Firefox/1.0
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.5) Gecko/20041107 Firefox/1.0

Can only give a clear example : try the site below in IE & Firefox and see the
difference for yourself.

http://www.bingoal.net/CONTENU/NLtableBEL1.shtml


Reproducible: Always

Steps to Reproduce:
1.Access the site with IE
2.Access the site with FireFox
3.See the difference

Comment 1

[Bogdan Stroe]

The HTML code of that page looks strange with tags like <DATEMA> and <RESULTAT>.
Maybe they use something specific to IE.

Comment 2

[Boris Zbarsky [:bzbarsky]]

> 1.Access the site with IE

No can do; no IE anywhere near here... can you post a screenshot of what IE shows?

Comment 3

[hmikron]

(In reply to comment #2)
> > 1.Access the site with IE
> 
> No can do; no IE anywhere near here... can you post a screenshot of what IE shows?

Here's how IE renders it : http://www.buffalo.easynet.be/Render0000.jpg

Comment 4

[Boris Zbarsky [:bzbarsky]]

Does IE get sent the same source?  There are </tr> tags in that page that IE
seems to be ignoring, if it renders the way that screenshot shows...

Comment 5

[Jason Barnabe (np)]

IE gets sent the same source.

There are 24 tr opening tags, and 123 tr closing tags...

Comment 6

[Dan]

(In reply to comment #5)
> IE gets sent the same source.
> 
> There are 24 tr opening tags, and 123 tr closing tags...

if you replace the extra </tr> tags (the ones that should be </td>) with
</td></tr> then IE renders it the same way as firefox... IE seems to be
interpreting what "should" have been written, rather than what is written.

Comment 7

[Boris Zbarsky [:bzbarsky]]

Dan, do you mean replacing "</tr>" with "</td></tr>", or do you mean replacing
"</tr></tr>" with "</td></tr>"?

Note that in HTML "</td>" is optional, so "</tr>" and "</td></tr>" are the same
thing.

Comment 8

[Dan]

I did not know that..., but yes, replacing the lines that are written as
<td class="CED">21.05  </tr>
with
<td class="CED">21.05  </td></tr>
causes IE to display the same as Firefox, "buggy", while
replacing the above line with
<td class="CED">21.05 </td>
or
<td class="CED">21.05
caused Firefox to display it the same way that IE does normally.  If you read
the source it looks like the extra </tr> tags are supposed to be </td> tags,
whoever wrote the table doesn't really want a new table row there, they just
want a new table data.  Firefox reads the </tr> tag and starts a new row, while
IE waits for the <tr> tag to actually start the new row.
  The table is poorly written, but if firefox waited for that <tr> tag to
actually start its new table row, then it wouldn't display poorly.

Comment 9

[Boris Zbarsky [:bzbarsky]]

Attachment: Minimalish testcase?

Dan, does IE render this with just three rows?

Comment 10

[Boris Zbarsky [:bzbarsky]]

In any case, clearly not a layout issue.  Blake, want to try to figure out what
IE's behavior wrt to optional table element end tags is, exactly, and whether
this one page is enough reason to try to duplicate its bugs?

Comment 11

[José Jeria]

(In reply to comment #9)
> Does IE render this with just three rows?

Yes

Comment 12

[Boris Zbarsky [:bzbarsky]]

Attachment: What about this testcase?

Comment 13

[José Jeria]

(In reply to comment #12)
> What about this testcase?

Same thing here, 3 rows ( http://www.jeria.net/mozilla/attachment_180590.gif )

Comment 14

[Boris Zbarsky [:bzbarsky]]

So it sounds like IE is just violating the HTML spec (which clearly says that
"</tr>" and "</td></tr>" are exactly equivalent.....)

Comment 15

[Dan]

Right, but we might also be violating HTML spec by starting a new table row
without a "<tr>" tag.  Of course, IE is continuing the same row, even though it
ended already, "</tr>".  Which is more important, ending rows at </tr> or only
starting rows at <tr>?

Either way we go, someone will mistype their html code...

Comment 16

[Boris Zbarsky [:bzbarsky]]

> we might also be violating HTML spec by starting a new table row without a
> "<tr>" tag. 

No, we're not.  The HTML spec explicitly says that any behavior is OK there
(that's the first point where the markup is invalid, and for invalid markup HTML
allows arbitrary behavior).

Comment 17

[Henri Sivonen (:hsivonen) (not reading bugmail until 2024-09-02)]

WONTFIX per HTML5. If you disagree, please file a bug against the HTML5 spec in the W3C's Bugzilla.